Fig 1.
Putative transport cycle of P-glycoprotein.
(A) The transport cycle is assumed to start from the inward-open conformation, followed by the occluded state assumed to form upon ATP and/or substrate binding. Finally, the transporter adopts the outward-open conformation driven by ATP hydrolysis and subsequently, the substrate is released in the extracellular space. (B) The inward-open conformation of crystallised mouse P-gp structure (4M1M) contains 12 transmembrane helices (TM1-12) and two nucleotide binding domains (NBDs). Helices TM1-6 and NBD1 are shown in light blue, while helices TM7-12 and NBD2 are given in dark blue shade. The conserved residues involved in ATP binding and hydrolysis, Walker A and signature motifs, are highlighted in orange and yellow, respectively (see S1 Fig for the back view). (C) Overlap of the three crystal structure models of mouse P-gp solved at 3.8 Å: 3G5U (yellow), 4KSB (green) and 4M1M (blue). Differences in the secondary structure assignment of TM helices between 3G5U and 4M1M are visible, as well as wider separation between NBDs found in 4KSB model.
Table 1.
Refinement statistics for the three P-glycoprotein crystallographic models [15–17].
Fig 2.
Pairwise RMSD and principal component analysis (PCA).
Heatmap representing pairwise RMSD [nm] calculated for the backbone atoms of P-gp conformations sampled in all 9 trajectories started from 3G5U, 4KSB and 4M1M for (A) the entire protein and (B) the TMD coordinate subset only. Black gridlines separate the three systems, while each replica (3G5U#1-3, 4KSB#1-3, 4M1M#1-3) is visible as a square formation around diagonal line (diagonal corresponds to the RMSD of a structure to itself, which is 0 nm). Conformational space sampled in each simulation as a function of the first three principal components (PC1, PC2, PC3) obtained from the analysis of the concatenated 9 trajectories for (C) the entire protein and (D) the TMD only, respectively. Each replica is shown in a different shade of yellow (3G5U), green (4KSB) or blue (4M1M). The TMD residues used in the pairwise RMSD and PC analysis were Tyr41-Ala358 (TMD1) and Leu684-Phe990 (TMD2).
Fig 3.
Overlap of the final P-gp conformations after 200 ns obtained from three replicas performed for each system. Each replica is coloured in a different shade of yellow for 3G5U (A), green for 4KSB (B), and blue for the 4M1M (C) system, where the lightest shade corresponds to replica #1 and the darkest to replica #3. The left and right panels show the TMD1 and the TMD2 at the front, respectively. The middle panel gives the side-view with lipids interacting with the TM4/6 (left side) and the TM10/12 (right side) portal. The POPC molecules are shown in different shades of violet and again, the lightest shade corresponds to replica #1. The protein is shown in cartoon representation and the TM helices are represented as cylinders that follow the helix axis (bendices) to highlight the change in helical geometry [55]. The POPC/cholesterol bilayer is depicted using the VDW sphere representation of phosphate atoms (grey).
Table 2.
Calculated average helical content (and standard deviation) from DSSP analysis of TMD1 (TM1-6) and TMD2 (TM7-12) for each replica of simulated systems 3G5U, 4KSB and 4M1M.
The amount of helical content was also averaged over all three simulations. TMD1 spans residues Trp44-Asp366, while TMD2 spans residues Trp704-Thr1011.
Fig 4.
Helical content of transmembrane domain.
The fraction of helical content in TMD1 (A) and TMD2 (B) as function of time for each replica started from 3G5U, 4KSB and 4M1M models calculated from the secondary structure analysis performed using the DSSP algorithm.
Fig 5.
(A-C) Distances measured between the centres of mass (COM) of the nucleotide binding domains during each simulation starting from 3G5U (A), 4KSB (B), and 4M1M (C); (D-F) Distribution of the inter-NBD distances obtained for each replica independently and for all three replicas (grey), starting from 3G5U (D), 4KSB (E), and 4M1M (F).
Fig 6.
(A) Orientation of P-gp molecules in the crystal lattice from which the 3G5U and 4M1M structural models were derived. The asymmetric unit contains two P-gp molecules (the NBDs highlighted in violet and magenta, respectively). (B) The crystal lattice representation of the 4KSB model, where the asymmetric unit contains a single P-gp molecule. Both crystal lattices belong to the P21 21 21 space group and the unit cell is represented with the blue wireframe.